test(sketchdb): HTTP e2e for capability-miss feedback loop

TODO.md

@@ -48,22 +48,34 @@ Deliverable: `AccuracyProfile::derive(&AggregationConfig)`
 returns concrete (ε, δ) per sketch type — not a stub. Unit tests
 against the sketchlib-bench corpus.
 
-### 3. End-to-end capability-miss feedback loop test
-
-The flow exists as code (`ControllerClient::create_plan` on
-backend side, `/api/v1/plan` route on controller side), but
-isn't tested end-to-end. For the paper's "controller reacts to
-workload drift" claim we need:
-
-- Test harness spins up controller + backend in-process (or
-  via Docker compose fixture)
-- Issues query that causes capability-miss
-- Asserts a new `StreamingConfig` arrives at the backend within
-  `N` seconds
-- Asserts next matching query hits (no miss)
-- Measures time-to-plan-ready + time-to-first-hit
-
-Not new engineering — testing the existing code path end-to-end.
+### 3. End-to-end capability-miss feedback loop test — **done (HTTP round-trip)**
+
+HTTP-level e2e landed in
+[`asap-query-engine/src/tests/capability_miss_http_e2e_tests.rs`](asap-query-engine/src/tests/capability_miss_http_e2e_tests.rs).
+Spins up a real backend HTTP server + mock controller HTTP
+server, fires a PromQL `sum(metric)` query that capability-misses,
+and measures wall-clock `time_to_plan_ready` from query issue to
+the backend observing the new `StreamingConfig` via
+`GET /api/v1/streaming-config`. Localhost floor: ~20 ms.
+
+Also asserts:
+- Mock controller received the HTTP notify with the documented
+  `{kind: "capability_miss", ...}` payload
+- Backend's hot-reload handle has the exact `agg_id` the
+  controller pushed
+- A repeat query on the same metric does **not** fire a second
+  notify (loop is idempotent under query replay)
+
+Follow-up (not paper-blocking):
+
+- **Cross-process test.** Today's test is single-process with
+  two HTTP servers. A docker-compose harness that wires a real
+  DataCollector controller binary against the real backend
+  binary is tracked by blocker #6 of `DataCollector/TODO.md`.
+- **Time-to-first-hit over real data.** The "next query returns
+  data" half of the story needs OTLP ingestion between the
+  plan push and the repeat query — still tracked as an
+  operational follow-up in the DataCollector repo.
 
 ### 4. Serialization format versioning tests